Local AI App Integration
amd/skills
Integrates local AI capabilities into applications using Embeddable Lemonade.
Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better.
Install Teach A Model
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install Aseiel/VideoHighlighter teach-a-model --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/teach-a-model .claude/skills/teach-a-model && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .claude/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-modelType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install Aseiel/VideoHighlighter teach-a-model --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/teach-a-model .agents/skills/teach-a-model && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .agents/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install Aseiel/VideoHighlighter teach-a-model --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/teach-a-model .cursor/skills/teach-a-model && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .cursor/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/Aseiel/VideoHighlighter.git --path .claude/skills/teach-a-model--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install Aseiel/VideoHighlighter teach-a-model --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/teach-a-model .gemini/skills/teach-a-model && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .gemini/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install Aseiel/VideoHighlighter teach-a-modelInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/teach-a-model .github/skills/teach-a-model && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .github/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install Aseiel/VideoHighlighter teach-a-model --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/Aseiel/VideoHighlighter.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/teach-a-model .opencode/skills/teach-a-model && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "teach-a-model" agent skill from https://github.com/Aseiel/VideoHighlighter/tree/main/.claude/skills/teach-a-model into .opencode/skills/teach-a-model/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "teach-a-model", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
teach-a-model- GitHub stars
- 160
- Token cost
- ~839 tokens
- SKILL.md length
- 469 words
- Files
- 1
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- AGPL-3.0
At a glance
Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better.
- Works in 3 steps: python -m modules.teach --project status → Run next.command, filling any from what… → Repeat. Stop and ask the user when
- The user asks to teach
- SKILL.md covers Loop, Fastest start, Starting a project and Reviewing (judge steps), plus 1 more section
- Calls python
What it does
Teach A Model is an agent skill from Aseiel/VideoHighlighter. Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better. Use when the user asks to teach, label, or train the app to recognise something new.
Its SKILL.md is about 840 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in AI & LLM Engineering, covering Computer vision, LLM inference and serving and Transcription. It works with Ollama and CUDA. The repository describes itself as: Open-source local AI video analyzer powered by Ollama. Visual search, automatic highlights, scene/action/object detection, audio analysis, and subtitle generation. Free, offline… The licence is AGPL-3.0.
When your agent uses it
- The user asks to teach
- Train the app to recognise something new
Example prompts
- “/teach-a-model”
Requirements
- Python 3
Workflow steps
3 steps, taken from the first numbered list in SKILL.md.
- python -m modules.teach --project status
- Run next.command, filling any from what the user said.
- Repeat. Stop and ask the user when
What it can do on your machine
Read from SKILL.md and the folder at commit deea999. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Shell commands in SKILL.md call:
python
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Teach A Model loads about 839 tokens when it runs. Until then it costs about 72 tokens; SKILL.md has 469 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
SKILL.md
The full file from Aseiel/VideoHighlighter at commit deea999, republished under its AGPL-3.0 licence (© Aseiel). 469 words, ~839 tokens.
Download SKILL.mdSave it as
.claude/skills/teach-a-model/SKILL.md (or your agent's skills folder).- name
- teach-a-model
- description
- Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better. Use when the user asks to teach, label, or train the app to recognise something new.
Teach a model
The pipeline is python -m modules.teach (docs/TEACH-A-MODEL.md). Every
command prints one JSON object; status names the next command.
Loop
python -m modules.teach --project <name> status- Run
next.command, filling any<placeholder>from what the user said. - Repeat. Stop and ask the user when:
- no classes exist and the user has not said what to find;
next.whoisjudgeand you cannot see images;- a command returns
error, or training fails (readruns/<n>/train.log).
Fastest start
Run doctor first on a machine you haven't used. If ready is false, relay
each blocking check's fix to the user rather than working around it.
If the user has example clips, ask them to put them in one subfolder per
class, named after what it shows, then run
quick --task <actions|objects> --examples <folder> --videos <files/folders>.
It runs every unattended step. After a judge step, auto continues
(auto --train includes training). A person reviews fastest with
review --window (and boxes review --window for boxes). Tiles show their
guesses; they click the wrong ones and press Enter.
Starting a project
- Ask what to find and for videos (files, folders or URLs) if not given.
init --task actionsfor something that happens over time (a movement, an activity);init --task objectsfor a thing visible in one frame.- Name classes with
check-namefirst. With example clips, runsuggest-names --clip ...: reuse afit: "good"stock label; otherwise choose a descriptive name in the same style, and always add--description. Ifsplitis not null, tell the user the examples look like two things and propose two classes. - Put the user's example clips in with
add-example: they make sorting far better than words alone.
Show full SKILL.md (190 more words)Show less
Reviewing (judge steps)
reviewreturnsimage: open it and look at every numbered tile. Tiles show start / middle / end of a clip; the caption is the guess.- Answer with
verdict --sheet Nusing--accept,--relabel N=<class>,--negative(none of the classes) and--reject(unusable). Decide every tile you can see. Use--accept-restonly after checking each unmentioned tile really matches its caption. - When unsure about a tile, reject it: a wrong label hurts more than a missing one.
- Run
sortagain after every one or two sheets (the window does it for you). - Tiles captioned "spot check" were auto-accepted. Judge them as strictly as any other tile: they are how auto-accept is kept honest, and training waits until each class has a few.
Rules
- Never edit
project.json/samples.jsonby hand; use the commands. - Never train with
--install alwaysunless the user asks: the default installs only a model that beats the installed one on the held-out set. - Report results as the per-class accuracy from the round's metrics
("finds
<class>in 7 of 10 held-out clips"), not as a single score. - Content-neutral: never add class names or presets to the repository.
© Aseiel, AGPL-3.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
Just SKILL.md in .claude/skills/teach-a-model of Aseiel/VideoHighlighter.
Open the folder on GitHubat commit deea999
Compare with similar skills
Teach A Model next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Teach A Model this skillAseiel/VideoHighlighter | 160 | — | ~839 | Automated safety check: Pass | AGPL-3.0 | |
| Local AI App Integrationamd/skills | 398 | — | ~6k | Automated safety check: Pass | MIT | |
| Llama Cppmagnus919/agent-skills | 113 | — | ~2.3k | Automated safety check: Pass | MIT | |
| Visiongridaco/grida | 2.7k | — | ~1.5k | Automated safety check: Pass | Apache-2.0 | |
| Deepstream DevNVIDIA/skills | 3.5k | — | ~3.3k | Automated safety check: Pass | Apache-2.0 | |
| AI SDK Developmenttrypostit/trypost | 678 | 2 repos | ~3.5k | Automated safety check: Pass | MIT |
Similar skills
Llama Cpp
magnus919/agent-skills
Operate, configure, benchmark, and troubleshoot llama.cpp across CPU, Metal, CUDA, HIP/ROCm, Vulkan, SYCL, and hybrid or multi-GPU systems.
Vision
gridaco/grida
Query images with a local Ollama vision model without loading the image into the main agent context.
- Official
Deepstream Dev
NVIDIA/skills
NVIDIA DeepStream SDK development with Python pyservicemaker API.
AI SDK Development
trypostit/trypost
TRIGGER when working with ai-sdk which is Laravel official first-party AI SDK.
Gemini Video Understanding
einverne/dotfiles
Analyze videos using Google's Gemini API - describe content, answer questions, transcribe audio with visual descriptions, reference timestamps, clip videos, and process YouTube URLs.
Related
Categories
Questions about Teach A Model
What does Teach A Model do?
Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better. Teach A Model is an agent skill from Aseiel/VideoHighlighter. Train a custom VideoHighlighter action or object model from a few videos the user provides — cut into samples, sort with CLIP, review contact sheets, build, train, install only if better.
When should I use Teach A Model?
Teach A Model fits situations like: the user asks to teach; train the app to recognise something new.
How do I install Teach A Model in Claude Code?
Run `npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a claude-code`. Or copy the skill folder (.claude/skills/teach-a-model in Aseiel/VideoHighlighter) into .claude/skills/teach-a-model in your project. Claude Code loads it when a task matches its description.
How do I install Teach A Model in Codex?
Run `npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a codex`. Or copy the skill folder (.claude/skills/teach-a-model in Aseiel/VideoHighlighter) into .agents/skills/teach-a-model in your project. Codex loads it when a task matches its description.
Can I use Teach A Model in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Aseiel/VideoHighlighter --skill teach-a-model -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/teach-a-model, .gemini/skills/teach-a-model, .github/skills/teach-a-model and .opencode/skills/teach-a-model in your project.
What does Teach A Model need to run?
Going by SKILL.md and its folder, Teach A Model needs the command-line tools its instructions call (python). Our summary lists: Python 3.
Does Teach A Model access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Teach A Model safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
What licence does Teach A Model use?
Teach A Model is published under the AGPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Teach A Model use?
About 839 tokens (SKILL.md is roughly 3.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Teach A Model?
Skills that share tags, products or a category with Teach A Model: Local AI App Integration (amd/skills, 398 stars), Llama Cpp (magnus919/agent-skills, 113 stars), Vision (gridaco/grida, 2.7k stars) and Deepstream Dev (NVIDIA/skills, 3.5k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Teach A Model?
Aseiel (a GitHub user) maintains it in Aseiel/VideoHighlighter, which has 160 GitHub stars. The repository was last updated on October 7, 2026.
Source: Aseiel/VideoHighlighter on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.